Yin-ting Lin¹ (), Andrew B. Leber¹; ¹The Ohio State University

Visual working memory is highly limited, and there are large individual differences in this limitation. Many studies suggest that individual differences in visual working memory are related to storage capacity and attentional control abilities. However, recent work has proposed that these individual differences may also reflect how efficiently individuals encode information into working memory (e.g., Nassar et al., 2018). Here we ask whether individuals with high visual working memory capacity are more likely to use chunking strategies to optimize encoding of information. In a continuous report task, participants viewed a display containing six colours and subsequently reported the colour of three items on a colour wheel. Critically, we showed participants the same sets of displays, and participants were free to choose which three items they report on each trial. This allows us to examine the degree to which participants selectively chunk and report similar items for each display. We also independently measured individuals’ working memory capacity in a change localization task. Here, we computed within-cluster variance in order to quantify the colour similarity for each display (i.e., the extent that the colours in the display can be chunked). Results show that memory performance is better for displays with high colour similarity, replicating previous work (Nassar et al., 2018). Furthermore, we assessed the extent that individuals use chunking strategies by examining the colour similarity of items that individuals chose to report on each trial. However, we found no evidence that individuals with high capacity were more likely to report items sharing greater colour similarity. Taken together, these findings suggest that while chunking strategies improve performance, individual differences in working memory do not seem to predict the use of chunking strategies. Future work will examine the impact of working memory load on the use of chunking strategies across individuals.

Acknowledgements: This work is supported by NSF BCS-2021038 to ABL